Previous Abstract Return to Session D3 Next Abstract

Session D3: GNSS Augmentation and Robustness for Autonomous Navigation (GNSS Integrity Augmentation)

Updated Airborne Multipath Models for Dual-Frequency Multi-Constellation GBAS
Maria Caamano, Daniel Gerbeth, Stefano Caizzone, German Aerospace Center, (DLR); Matteo Sgammini, European Commission, Joint Research Centre (JRC)
Date/Time: Thursday, Sep. 19, 11:26 a.m.

Best Presentation

Currently proposed airborne multipath models are based on 100-second time-invariant (for single-frequency L1/E1 and L5/E5a signals) and time-variant (for the ionospheric-free combination of these signals) smoothing to describe the behavior of the aiborne multipath error in different applications (e.g. Satellite Based Augmentation Systems and Advance Receiver Autonomous Integrity Monitoring). However, the recently proposed architecture for dual-frequency multi-constellation (DFMC) GroundBased Augmentation Systems (GBAS) incorporates different smoothing characteristics, as the use of divergence-free time-variant smoothing also for L1/E1 and L5/E5a with time constants ranging from raw data up to 600 seconds.
Therefore, this paper presents extended airborne multipath models for single-frequency GPS and Galileo L1/E1, L5/E5a, and the ionospheric-free combination to support standardization of DFMC GBAS. To derive these models, data collected during dedicated flight campaigns within the European project DUFMAN (Dual Frequency Multipath Model for Aviation) has been reprocessed using a methodology similar to that employed in the DUFMAN project itself. This methodology involves estimating multipath from divergence-free code-minus-carrier observables, removing cycle slips, calibrating airborne antenna errors, and resolving ambiguities. Additionally, the behavior of estimated multipath during the transient phase of the time-variant filter is studied.
The results include proposed models for both unsmoothed and 600-second smoothed multipath after filter convergence, as well as inflation factors to adjust the proposed 600-second model to account for increased errors during the filter’s transient phase.

Previous Abstract Return to Session D3 Next Abstract